Every time I see a BOM discussion on LinkedIn, my mind automatically returns to a specific role I had managing the BOM at a manufacturing startup. At the time, we were working in Excel, manually digging through CAD files stored on a shared drive.

Initially, I worked in Manufacturing Engineering, attempting to assign parts to processes. However, it quickly became apparent that we were ordering unnecessary parts, resulting in a growing pile of costly scrap. I eventually transferred to Supply Chain, where I realized that sourcing is the “magic” by which the BOM becomes real. My daughter was obsessed with the children’s book The Velveteen Rabbit, and at the time I realized that sourcing is where the rubber meets the road and parts become actual physical incarnations.

That experience was many years ago, but the analogy has stuck with me. I entered the PLM ecosystem from a deep parametric CAD background, and for a long time, it was clear that ERP was not architected to manage PLM. However, I have seen developments in the last few years that are changing my mind. Modern ERP platforms are becoming increasingly flexible and—for better or worse—are brilliantly suited to manage an “agentic army,” as they already manage a company’s human resources.

• SAP: In recent years, SAP has shifted PLM capabilities into the SAP Digital Supply Chain portfolio, heavily integrating them with S/4HANA.
• Oracle: Oracle Fusion Cloud PLM is built on a common data model that integrates directly with Oracle’s ERP and SCM (Supply Chain Management) suites, allowing capabilities to flow directly into manufacturing. If you hasn’t checked out Oracle’s AI Agent studio yet, you will want to.
• Bluestar PLM: Developed by the PDM Technology Group, this is natively embedded within the Microsoft Dynamics 365 ecosystem.

I know the idea of a single ecosystem to manage everything is often just a fantasy. As an old-school CAD geek, I realize that locally installed parametric CAD relies on intermingled hierarchies of drawings and assemblies. These often reference each other in a fashion resembling a Rube Goldberg machine—a problem much better solved on the PDM (Product Data Management) side.

If you aren’t familiar with Rube Goldberg machines, here are two of my favorites:

• OK Go: https://youtu.be/qybUFnY7Y8w
• Honda: https://youtu.be/_ve4M4UsJQo

My take is that AI may drive PLM functionality further into ERP, allowing the product to become “real” much earlier in the process, but PDM is here to stay.